The n e w e ng l a n d j o u r na l of m e dic i n e

Review Article

Dan L. Longo, M.D., Editor

Rheumatoid Arthritis — Common Origins,

Divergent Mechanisms
Ellen M. Gravallese, M.D., and Gary S. Firestein, M.D.​​

R
heumatoid arthritis is one of the most common immune-medi- From the Division of Rheumatology, In-
ated diseases. Its primary manifestation is inflammatory arthritis charac- flammation, and Immunity, Brigham and
Women’s Hospital, Boston (E.M.G.); and
terized by symmetric, polyarticular pain and swelling, typically involving the Division of Rheumatology, Allergy,
the small joints of the hands and feet. However, rheumatoid arthritis is a sys- and Immunology, University of California
temic disease associated with multiple coexisting conditions and extraarticular at San Diego School of Medicine, La Jolla
(G.S.F.). Dr. Gravallese can be contacted
manifestations. Onset of inflammatory synovitis results from the interactions of at ­egravallese@​­bwh​.­harvard​.­edu.
genetic factors and specific environmental exposures. The disease process begins
N Engl J Med 2023;388:529-42.
years before clinically apparent arthritis and manifests as a continuum that origi- DOI: 10.1056/NEJMra2103726
nates with asymptomatic immune dysfunction and progresses through various Copyright © 2023 Massachusetts Medical Society.
stages before the disease can be classified as rheumatoid arthritis.
This review focuses on seropositive rheumatoid arthritis, marked by the pres- CME
at NEJM.org
ence of autoantibodies to post-translationally modified proteins, including anti–
citrullinated protein antibodies (ACPAs, measured as anti–cyclic citrullinated
peptide antibodies); less specific autoantibodies, known as rheumatoid factors,
that bind the Fc portion of immunoglobulin; or both antibody types. Seronegative
rheumatoid arthritis is a separate entity marked by polyarthritis but with poorly
defined pathogenetic mechanisms. The course of seronegative rheumatoid arthri-
tis is typically less destructive to joints,1 but the approach to treatment is similar
to that of seropositive disease.
In contrast to an immune disease such as psoriasis, which largely depends on
the dominant interleukin-23–interleukin-17 pathway, rheumatoid arthritis has
multiple potential paths to a common clinical presentation. The disease progress-
es from preclinical rheumatoid arthritis through chronic disease and involves
pathogenic pathways and cell lineages that can differ among patients, complicat-
ing therapeutic efforts. The predominance of certain pathways over others in in-
dividual patients is underscored by the diversity of clinical responses to targeted
therapies, despite a remarkably similar clinical phenotype. There have been revolu-
tionary changes in the treatment of rheumatoid arthritis in the past three decades,
but many patients still have persistent disease. The ability to identify the specific
pathogenic mechanisms in individual patients would improve outcomes by direct-
ing therapy to those targets.
The preclinical stages of seropositive rheumatoid arthritis are characterized by
disordered immunity, often associated with mucosal surfaces, including the oral
cavity, lungs, and gastrointestinal tract, and by local and systemic generation of
ACPAs. These autoantibodies can be detected in the blood a median of 4.5 years
before the onset of arthritis.2 The risk of rheumatoid arthritis increases over time
as autoantibody levels increase. As this preclinical phase progresses, ACPAs directed
against an expanding array of protein epitopes ensue, along with a rise in pro­
inflammatory proteins in blood, ultimately resulting in joint inflammation.3
Immune responses to altered peptides are not limited to citrullination; carbamy­la­
tion, malon­dialdehyde–acetaldehyde adduct formation, and other protein modifi-

n engl j med 388;6  nejm.org  February 9, 2023 529

The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Risk factors
• Genetics
• Environmental
exposures Potential
(including smoking) treatment Cell lineage or preclinical
for disease driver-specific mechanisms
• Mucosal
inflammation prevention • Synovial inflammation
• Protein modification • Antigen-driven T- and
(e.g., citrullination) • Second hit B-cell immunity
• Autoantibody • Multiple possible • Fibroblast activation Window of opportunity for
production mechanisms • Epigenetic remodeling treatment to prevent chronicity

Preclinical RA RA Early RA Established

Healthy RA
At risk for RA Transition Clinical synovitis
Personalized mechanisms:
• Immune activation • T cells • New adaptive T-cell Joint damage
and loss of • B cells subsets
Osteoclasts
tolerance • Fibroblasts • Synovial and proteases
• Increased • Inflammatory cytokines autoantibody
autoantibody levels production RA flare
• Other cell types
and affinity • New fibroblast Preinflammatory
• Epigenetic subpopulations and mesenchymal
remodeling aggressive behavior (PRIME) cells

Figure 1. Initiation and Progression of Rheumatoid Arthritis (RA).

The boxes show various stages of the disease continuum and the pathogenic features of each stage. The purple text shows characteristics
that affect the transition between disease stages. RA progresses from a healthy state to preclinical RA (at risk for RA) to the RA transition
to early synovitis and finally to established, destructive disease. The pathway is not unidirectional, since persons in the disease stage
before synovitis who are positive for antibodies against citrullinated peptides (ACPAs) can become ACPA-negative, and in some ACPA-
positive persons, disease never develops. The continuum of disease evolution offers potential opportunities for the prevention of RA.
Although each disease state has a characteristic clinical phenotype, multiple pathways and mechanisms can contribute to pathogenesis
for an individual patient. This is depicted by the red text and arrows, which indicate disease that is dependent on a particular cell type
or mediator. Therapeutic approaches should ideally be targeted to address the particular pathogenic mechanism in an individual patient.
Some patients may have disease that is characterized by multiple mechanisms, resulting in a partial response or a lack of response to a
given targeted therapy.

cations can induce recognition of neoantigens, 1.0%, although the prevalence is higher in cer-
with the production of antibodies to these modi-
tain populations, such as Indigenous North
fied protein antigens.4 Americans. Rheumatoid arthritis can occur at
Treatments are designed to induce clinical any age, but the incidence peaks in the third
remission in patients with established rheuma- through fifth decades of life, and the disease is
toid arthritis. In addition, disease prevention2 to 3 times as common among women as it is
strategies are being developed for persons con-among men. The effects of estrogen on immune
sidered to be at risk for the disease on the basis
function probably play a part in the female pre-
of family history, autoantibody status, geneticdominance of the disease,5 although additional
risk factors, or a combination of these find- sex-related factors are also likely to be involved.
ings, as well as for persons with very early Several infectious agents have been proposed
stages of joint pain or inflammation, before as etiologic or contributing agents, including
rheumatoid arthritis has been definitively diag-
Epstein–Barr virus, retroviruses, bacterial super-
nosed (Fig. 1). antigens, and mycoplasma species, as well as
organisms such as oral Porphyromonas gingivalis
and gut prevotella species.6,7 However, a single
Epidemiol o gy a nd Dise a se
Cl a ssific at ion microorganism that accounts for all patients is
unlikely to be causal. The most prominent be-
The prevalence of rheumatoid arthritis is re- havioral risk factor for the development of rheu-
markably consistent worldwide, at about 0.5 to matoid arthritis is cigarette smoking. Additional

530 n engl j med 388;6 nejm.org February 9, 2023

The New England Journal of Medicine

Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
 Rheumatoid Arthritis
factors marginally increase the risk of rheuma-
toid arthritis, including obesity, low vitamin D
levels, and use of oral contraceptives. Factors
that decrease the risk include a Mediterranean
diet, n–3 fatty acid intake, fish oil supplementa-
tion, and alcohol consumption.8,9
Although rheumatoid arthritis is characteris-
tically marked by symmetric arthritis in the
small joints of the hands and feet, as the disease
progresses, any synovial joint can be involved.
The 2010 American College of Rheumatology–
European League against Rheumatism classifi-
cation criteria10 focus on earlier disease manifes-
tations than did previous classification criteria,
with the introduction of a composite scoring
system that includes the number and site of
clinically involved joints, the duration of symp-
toms, and the status with respect to rheumatoid
factor, ACPAs, and acute-phase reactants. ACPAs
are increasingly used to support the diagnosis
because of their high specificity.
Gene t ic R isk a nd Epigene t ic
Fac t or s
The most prominent risk factor for rheumatoid
arthritis is genetic. For first-degree relatives of
patients with rheumatoid arthritis, the risk of
disease is increased by a factor of 2 to 5. The
HLA-DR locus is the most important genetic
association. Well-characterized sequences in the
hypervariable region of the HLA-DRβ chain
(amino acids 70–74), known as the “shared epi-
tope,” are associated with an increased risk.
HLA-DR is involved in antigen presentation to
CD4+ T cells and could increase susceptibility
through its ability to bind and present specific
arthritogenic peptides. HLA-DR genes associated
with rheumatoid arthritis can bind peptides mod-
ified by citrullination more avidly than native
peptides, inducing T-cell activation and cytokine
production. In addition, these HLA molecules
may influence T-cell receptor selection toward a
more autoimmune repertoire.11 Informatics analy-
sis of major histocompatibility complex (MHC)
data indicates that three amino acid positions
in HLA-DRβ1 and a single amino acid in HLA-B
and HLA-DPβ1 that modify the peptide-binding
groove explain most of the MHC association
with disease risk.12 HLA-DRB1 is associated not
only with susceptibility but also with disease
severity and possibly with varying treatment
responses to certain biologic agents.13,14
n engl j med 388;6  nejm.org  February 9, 2023
The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
More than 100 additional alleles have been
identified that contribute to the risk of disease
and overwhelmingly implicate immune pathways.
Many are located in gene regulatory or intronic
regions, but some involve the coding region and
affect gene function. For example, a polymor-
phism in PTPN22, a phosphatase involved in T-cell
receptor signaling, is one of the best-character-
ized alleles associated with rheumatoid arthritis.
R620W, a gain-of-function amino acid change in
PTPN22, increases disease risk by a factor of
more than 2.15 Many other risk alleles are also
associated with immune processes, including the
coding region of the interleukin-6 receptor and
noncoding regions near the TRAF1–C5 locus.
Most of these alleles marginally increase the
odds ratio for rheumatoid arthritis, by a factor
of approximately 1.1 to 1.2.
The relatively low concordance of rheumatoid
arthritis in monozygotic twins (approximately
15%), as compared with the concordance of mono-
genic diseases, suggests that noncoding DNA
epigenetic marks, possibly induced by environ-
mental or stochastic factors, are also important.
DNA methylation might contribute to disease sus-
ceptibility, as suggested by distinct methylation
patterns in twins who are discordant for rheu-
matoid arthritis.16 Furthermore, in at-risk persons
without synovitis who have high blood levels of
rheumatoid factor or ACPAs, peripheral-blood
mononuclear cells are characterized by abnormal
DNA methylation in immune-related genes years
before the onset of symptoms.17 Later, T cells with
aberrant epigenetic marks in immunologic path-
ways accumulate in the inflamed synovium.18 In
contrast, patients with osteoarthritis have fewer
differentially marked genes in synovial T cells,
and they are randomly distributed. Thus, remodel-
ing of the disease-associated epigenome in
synovium could be driven by processes that con-
tribute to the transition from preclinical to
clinical rheumatoid arthritis.
From Muc os a l Infl a m m at ion
t o A lter ed Pep t ide s t o Cl inic a l
Dise a se
Environmental and behavioral influences play a
major role in susceptibility to rheumatoid arthri-
tis and disease severity. Cigarette smoking and
genetic risk can be synergistic: for ACPA-positive
smokers with two copies of the susceptibility
shared epitope, the risk of rheumatoid arthritis
531
 The n e w e ng l a n d j o u r na l
is 20 times that for nonsmokers.19 The risk gradu-
ally abates after smoking cessation, approaching
the risk for nonsmokers within two to three
decades.20 Inflammation and stress at mucosal
surfaces, induced by environmental exposures
such as cigarette smoke, contribute to disease
initiation in persons with risk alleles for rheu-
matoid arthritis, and the link between mucosal
inflammation and rheumatoid arthritis is stron-
gest for the airway. The mammalian genome in-
cludes enzymes known as peptidyl arginine
deiminases, which convert arginine to citrulline.
These enzymes are induced by cell stress and
lead to post-translational citrullination of many
proteins. Citrullination is quite active in the air-
ways of smokers, where modified peptides have
been detected in macrophages.21
Bronchiolar thickening and local neutrophil
extracellular trap formation occur in asymptom-
atic at-risk persons with high ACPA titers, as well
as in first-degree relatives of patients with rheu-
matoid arthritis.22 The extruded DNA in neutro-
phil extracellular traps forms a scaffold for citrul-
linated peptides and amplifies immune responses
that can generate ACPAs.23 In at-risk persons, the
combination of peptide citrullination and HLA-
DR haplotypes that bind citrullinated peptides
more avidly than native peptides24 can lead to a
local immune response and further ACPA produc-
tion. However, ACPAs produced at sites of muco-
sal damage might also serve as a mechanism to
clear citrullinated proteins.25 The evolving ACPAs
arise from B cells and plasmablasts through af-
finity maturation, which is driven by specific
citrullinated proteins and oligoclonal expansion of
antigen-specific cells.26 Concomitant increases
in serum cytokines and chemokines provide evi-
dence of early systemic inflammation that ulti-
mately culminates in symptomatic joint inflam-
mation.
Production of ACPAs and other autoantibod-
ies represents a break in tolerance. Such breaks
can be facilitated by the selective introduction of
N-linked glycosylation sites in the B-cell recep-
tor antigen-binding pocket, which alters the anti-
gen-binding site and enhances B-cell activation.27
ACPAs (IgA or IgG) can bind to an array of
citrullinated proteins, including fibronectin,
enolase, histones, and fibrinogen. The mere
presence of ACPAs is not sufficient to induce
arthritis, and a consistent pattern of citrullinat-
ed proteins or antibody levels that precede or
532 n engl j med 388;6  nejm.org  February 9, 2023
The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
coincide with synovitis has not been identified
in patients with rheumatoid arthritis. Synovial-
biopsy specimens from persons with preclinical
rheumatoid arthritis and arthralgias show little
or no evidence of inflammation or local immune
responses despite high levels of circulating
ACPAs.28 In preclinical models, the administra-
tion of ACPA autoantibodies does not cause ar-
thritis29 but can exacerbate existing synovitis.
Even so, rising ACPA titers in humans are a
harbinger of clinical disease. When the titers in
at-risk persons reach 3 times the upper limit of
the normal range, there is a 30 to 50% chance
that rheumatoid arthritis will be diagnosed
within 3 to 5 years.30
These observations raise the possibility that
progression from preclinical rheumatoid arthri-
tis to established disease might be prevented
through therapeutic intervention or mitigation
of environmental stress. Several clinical trials un-
successfully attempted to intercede in this tran-
sition, including treatment with atorvastatin31
and B-cell depletion with a single course of
rituximab.32 The latter delayed, but did not pre-
vent, conversion to clinical rheumatoid arthritis
in ACPA-positive persons presenting with arthral-
gias. In at-risk persons with arthralgias and
imaging evidence of synovitis, 1 year of treat-
ment with methotrexate, a nonadaptive immune-
system intervention, did not prevent rheumatoid
arthritis, as assessed after 2 years. However, the
disease was less severe in the treated cohort.33
Although proinflammatory processes have been
emphasized in the transition to clinical disease,
inadequate production of antiinflammatory cyto-
kines, such as interleukin-1 receptor antagonist
(interleukin-1Ra) and interleukin-10, or defective
synovial apoptosis34 could also contribute to the
onset and perpetuation of disease.
He tero genei t y of S y nov i t is
in R heum at oid A r thr i t is
Synovitis is a hallmark of rheumatoid arthritis,
with an influx of inflammatory cells leading to
multiple villous projections within the joint cav-
ity. Typical histologic features include synovial
hyperplasia, neovascularization, and a heteroge-
neous inflammatory infiltrate that can include
lymphoid aggregates and germinal center–like
structures. Infiltrating cells include T and B cells,
plasma cells, plasmablasts, macrophages, den-
 Rheumatoid Arthritis
dritic cells, and occasional mast cells and natu-
ral killer cells. Neutrophils are sparse in rheuma-
toid synovium but pass through tissues into
synovial fluid rapidly.
Synovial histologic and transcriptional analy-
ses show marked heterogeneity among patients
with established rheumatoid arthritis,35 perhaps
providing clues to pathogenic pathways that are
active in a given patient. Synovial assessments,
however, can be complicated by sampling bias
and by distinct epigenetic marks and transcrip-
tomes that depend on joint location.36 For exam-
ple, fibroblasts derived from hip and knee syno-
via can be distinguished from each other on the
basis of DNA methylation and transcriptome
patterns. Noncoding RNAs also vary according
to joint location and could help shape stromal
cell phenotypes and function.37 Although classifi-
cation systems have been proposed on the basis
of histologic features or the most prominent cell
types in a given patient’s synovial tissue on bi-
opsy,38 meaningful correlations between histo-
pathological patterns and clinical disease activ-
ity or outcomes are thus far limited. Tissue
analyses performed with RNA sequencing meth-
ods and stratification by cell lineage signatures
may have the potential for predicting the response
to a given therapy.39 Data from synovial biopsies
in patients with rheumatoid arthritis suggest
that molecular and histologic profiling might
provide insights into the response to the B-cell–
directed agent rituximab as compared with block-
ade of the interleukin-6 receptor with tocilizu­
mab.40 These approaches continue to be promising
research tools but have not yet defined specific
pathways driving disease in a given patient.
Approaches based on systems biology might
also help stratify patients according to shared
pathogenic pathways. A recent study integrated
transcription factor–binding site accessibility
with the transcriptome in fibroblasts.41 At least
two clusters of patients were identified on the
basis of divergent transcription factor functions
in cultured fibroblast-like synoviocytes. For ex-
ample, the transcription factor retinoic acid re-
ceptor alpha had proproliferative effects in the
transforming growth factor β pathway in one
cluster but antiproliferative effects in the other
cluster. Similar unbiased systems approaches
could provide insight into how biologic features
vary among patients with similar clinical pheno-
types.
n engl j med 388;6  nejm.org  February 9, 2023
The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
Ne w Insigh t s in t o Patho gene sis
Studies of specimens from ultrasound-guided
synovial biopsy in patients with rheumatoid ar-
thritis offer new insights into pathogenesis.
Evaluation of synovial cell surface markers with
cytometry by time-of-flight and single-cell RNA
sequencing has provided data on the large array
of cell lineages in rheumatoid synovium, includ-
ing more than 20 transcription-defined T-cell
subtypes.42 T cells are central in the pathogene-
sis of rheumatoid arthritis, and one important
remaining question is whether some of these
cell phenotypes are responsible for the disease,
represent a response to the synovial microenvi-
ronment, or are merely “spectators at a fire,”
recruited by the rich chemoattractant milieu.
Several novel and important T-cell subsets have
been identified, however, including peripheral
helper T (Tph) cells, located within synovial B-cell
clusters and in the circulation, which promote
B-cell production of interleukin-21, supporting
immunoglobulin affinity maturation, among
other functions (Table 1). Tph cells also promote
B-cell proliferation and differentiation into anti-
body-producing plasma cells.43 Oligoclonal ex-
pansion of synovial B cells with somatic muta-
tions, indicating local ACPA affinity maturation,
is also prominent in rheumatoid synovium.26
Studies using fate mapping systems have
identified novel macrophage subsets, including
CX3CR1+ tissue-resident macrophages that form
an immunologic barrier on the synovial surface,
restricting the flux of proteins across the normal
synovial lining.48 Identification of additional
macrophage subsets, including resident macro-
phages47 with an antiinflammatory phenotype
and inflammatory macrophages that contribute
to the production of proinflammatory factors,
also provides insights into pathogenesis. In ad-
dition, dendritic cells, which are present in rheu-
matoid synovium, play a part in local antigen pre-
sentation and activation of autoreactive T cells.56
Extensive work with synovial biopsy specimens
from patients with rheumatoid arthritis has
highlighted the importance of several novel fibro-
blast phenotypes that promote inflammation,
including those with proinflammatory functions
whose differentiation is regulated by endothelial
cells.51 In addition, the transcription factor ETS1
defines a fibroblast phenotype that regulates
bone damage through the production of RANKL
533
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 1. Recently Identified Cell Lineages and Cell Phenotypes Contributing to Rheumatoid Arthritis (RA).*

Cell Class Comments Study

T cells
CD4+ PD1+ CXCR5− (Tph) Located in lymphoid aggregates adjacent to B cells; produce interleukin-21, Rao et al.43
which supports B-cell proliferation and differentiation into plasma cells
CD8+ GZMK+ Located in RA synovial sublining; source of interferon-γ Jonsson et al.44
B cells
Mucosal, circulating, and Oligoclonal expansion with evidence of recirculation and ACPA affinity matura- Kongpachith et al.26
synovial B cells tion due to somatic mutation
NR4A+ Produces lymphotoxins and interleukin-6, which promote lymphoid aggregate Meednu et al.45
formation
Macrophages
MerTK− Proinflammatory; found in active RA synovium; associated with interleukin-6 Alivernini et al.46
and TNF production
MerTK+ Antiinflammatory; associated with lipoxin and resolvin production Cai et al.47
CX3CR1+ Resident cells in RA synovial lining with tight junctions that serve as immuno- Culemann et al.48
logic barrier; disrupted in RA synovium
Alternatively activated Interleukin-33 reprogrammed with metabolic rewiring that uncouples respira- Faas et al.49
tory chain and enhances inflammation resolution
HBEGF+ Promotes fibroblast aggressiveness and invasion Kuo et al.50
Fibroblasts
FAP+ CD90+ Proinflammatory phenotype located in sublining; regulated by NOTCH3 Wei et al.51
FAP+ CD90− Located in intimal lining; produce interleukin-6, metalloproteinases, and pros- Mizoguchi et al.52
tanoids in RA
ETS1 Regulates bone damage through production of RANKL by fibroblasts in synovium Yan et al.53
PRIME cells Present in circulation of patients with RA (according to transcriptome profile); Orange et al.54
increase in peripheral-blood PRIME cells precedes RA flares, possibly asso­
ciated with B-cell activation
Neutrophils: synovium and lung Produce NETs, which bind citrullinated peptides to enhance ACPA production Corsiero et al.55
mucosa

* ACPA denotes anti–citrullinated protein antibody, NETs neutrophil extracellular traps, PRIME preinflammatory mesenchymal, RANKL recep-
tor activator of nuclear factor-κB ligand, TNF tumor necrosis factor, and Tph peripheral helper T.

(receptor activator of nuclear factor-κB ligand) L ong -Ter m C onsequence s

and activation of osteoclasts.53
Murine studies indicate that fibroblasts have
the potential to migrate from an inflamed joint
through the bloodstream and could thus “spread”
synovitis.57 More recent studies in humans have
identified the appearance of preinflammatory
mesenchymal (PRIME) cells in peripheral blood
obtained just before disease flares in patients
with rheumatoid arthritis.54 Longitudinal tran-
scriptomic data also provide evidence of periph-
eral-blood B-cell activation before the appear-
ance of PRIME cells and disease flares.
Interactions between B cells and PRIME cells
could therefore serve as a potential target to
abrogate disease flares.
of Dise a se
Morbidity and Mortality
Patients with rheumatoid arthritis have an in-
creased risk of death. Cardiovascular disease is
the most common cause of premature death,
and the excess risk of cardiovascular disease in
rheumatoid arthritis is attributed to the combi-
nation of chronic inflammation and well-docu-
mented risk factors for cardiovascular disease
such as hypertension and dyslipidemia. Tradi-
tional risk factors (e.g., elevated low-density lipo-
protein cholesterol levels) have a weaker correla-
tion with cardiovascular risk in rheumatoid
arthritis than in the general population.58 Recent

534 n engl j med 388;6  nejm.org  February 9, 2023

The New England Journal of Medicine

Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
 Rheumatoid Arthritis

Figure 2. Complications of RA and Coexisting Conditions.

Proinflammatory cytokines, complement activation, Stroke
and immune complex formation in RA promote sys- Depression, mental
temic inflammation that affects many systems and re- health disorders
sults in complications and coexisting conditions, along
with the production of acute-phase reactants (APRs) Scleritis, episcleritis
mediated in large part by interleukin-6. Designation as Sicca syndrome
a complication or coexisting condition is not exact,
and the two categories can overlap. Altered lipid me- Bronchial inflammation,
tabolism and cytokines, including tumor necrosis fac- bronchiectasis
tor and interleukin-6, contribute to atherogenesis, Interstitial lung disease
myocardial infarction, and stroke. Pulmonary compli-
cations occur through multiple mechanisms. Rheuma- Myocardial infarction
toid nodules and vasculitis are seen less commonly Congestive heart failure
Valvular abnormalities
with current treatment approaches. Chronic inflamma-
Pericarditis
tion contributes to metabolic and other coexisting con-
ditions and can result in depression and altered coping
behaviors. Possible mechanisms affecting the immune Lymphoproliferative
disease
status of the central nervous system (CNS) include
effects of proinflammatory cytokines on activation of
Lipid abnormalities,
blood–brain barrier endothelium, leading to the trans- APRs
port of cytokines into the CNS, and effects on neural
circuits and plasticity.62 Proinflammatory cytokines
Vasculitis
synergize with RANKL (receptor activator of nuclear Nodules
factor-κB ligand) to promote osteoclastogenesis and
articular and systemic bone loss, leading to fractures,
Obesity
and inhibitors of the Wnt signaling pathway prevent
bone formation and erosion repair by osteoblasts. Joint
destruction is mitigated by tight control of inflamma- Joint damage
tion. The inflamed synovium produces multiple algo-
gens that increase pain sensitivity by reducing the fir- Low bone mineral
ing threshold for local nociceptors. Sensitization of density, osteoporosis,
fracture
central pain pathways by proinflammatory mediators
has also been implicated.63 The majority of patients
Pain
with RA have marked fatigue due to pain, sleep distur-
Functional deterioration
bance, and other factors.64 These complications and
Infection
coexisting conditions combine to result in a multifac-
torial deterioration of function over time. Susceptibility
to infections is increased in RA owing to impaired host
defense, and this can be exacerbated by the use of im- and progress in some patients, and are associ-
munosuppressive agents. ated with a higher risk of death from respiratory
disease.65 A variant in the promoter region of the
gene encoding mucin 5B (MUC5B) is associated
studies have shown coronary microvascular dys- with idiopathic pulmonary fibrosis, and this vari-
function in patients with rheumatoid arthritis, ant is also associated with the usual interstitial
like that seen in patients with diabetes,59 which pneumonia form of interstitial lung disease in
probably contributes to excess cardiovascular patients with rheumatoid arthritis,66 implicating
mortality. Current treatment strategies that re- mucins in the pathogenesis of this complication.
duce inflammation mitigate the risk of death.
The benefit of this approach is particularly well Joint Destruction
documented with tumor necrosis factor (TNF) The rheumatoid synovium is characterized by
blockers.60 expansion of tissue at the interface with carti-
Complications and coexisting conditions in lage and bone. This expanding tissue, known as
patients with rheumatoid arthritis, including an pannus, resembles a locally invasive tumor and
increased risk of lymphoproliferative disease,61 extends over the surface of cartilage. Pannus
are shown in Figure 2. As rheumatoid arthritis also invades the bone marrow space directly or
progresses, pulmonary interstitial fibrosis, bron- through pores in cortical bone. In active rheu-
chial inflammation, and bronchiectasis develop matoid arthritis, the extracellular matrix of

n engl j med 388;6 nejm.org February 9, 2023 535

The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
 536
Table 2. Agents Approved and in General Use or Not Approved for Use in RA.*

Class and Agent Mechanism of Action Comments

Approved and in general use
Traditional DMARDs
Methotrexate Affects multiple cell types; inhibits several pathways, including adenosine Used as a single agent or in combination ther­
­metabolism apy; most common initial agent
The

Leflunomide Inhibits dihydroorotate dehydrogenase and pyrimidine metabolism and may Used as a single agent or in combination
inhibit expansion of activated leukocytes therapy
Sulfasalazine Combination drug (5-aminosalicylic acid and sulfapyridine) that potentially inhibits Often used in combination with methotrexate
inflammatory cytokines and chemokines and alters adenosine metabolism and hydroxychloroquine
Hydroxychloroquine Possibly stabilizes macrophage lysosomes; modulates TLR7 and TLR9 activity Used as monotherapy for mild disease; often
used in combination with methotrexate
and sulfasalazine (triple therapy)
Biologic DMARDs
Cytokine inhibition Interruption of cytokine networks Often used in combination with methotrexate
or another traditional DMARD
TNF: adalimumab, certolizumab pegol, Blockade of TNF inhibits activation of leukocytes, FLS, endothelial cells, and os-
n e w e ng l a n d j o u r na l

etanercept, golimumab, infliximab teoclasts, preventing matrix degradation and production of proinflammatory

The New England Journal of Medicine

of

molecules
Interleukin-6: sarilumab, tocilizumab Blockade inhibits B-cell differentiation; activation of leukocytes, osteoclasts, and
acute-phase reactant elevation; lipid alterations

n engl j med 388;6  nejm.org  February 9, 2023

Interleukin-1: anakinra Interleukin-1Ra blocks interleukin-1 binding to receptor; inhibits activation of leuko- Less effective than other anticytokine agents

Copyright © 2023 Massachusetts Medical Society. All rights reserved.

cytes, FLS, endothelial cells, and osteoclasts, preventing matrix degradation in RA
m e dic i n e

T cells: abatacept Binds CD80 and CD86 and blocks T-cell costimulation, inhibiting naive T-cell Increased efficacy when used in combination
­activation with methotrexate
B cells: rituximab Binds CD20 and depletes B cells, inhibiting antigen presentation and autoantibody Often used in combination with methotrexate
production
Synthetic DMARDs–JAK inhibitors: baricitinib, Interrupt cytokine networks through blockade of JAK–STAT pathway, inhibiting FLS
tofacitinib, upadacitinib activation, leukocyte maturation, and autoantibody production

Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Selected biologic DMARD targets tested but
agent not approved
Cytokine inhibition Interruption of cytokine networks
 Class and Agent Mechanism of Action Comments
GM-CSF Blockade inhibits differentiation and activation of myeloid cells and granulocytes Efficacy in phase 2 studies; phase 3 studies in
progress
Interleukin-15 Blockade inhibits T-cell–FLS and T-cell–macrophage interactions and TNF Modest efficacy in clinical trials
­production
Interleukin-17A Blockade inhibits synergism with TNF and activation of FLS, osteoclasts, and Modest efficacy in phase 3 trials
­chondrocytes
Interleukin-18 Blockade inhibits granulocyte and osteoclast activation and Th1 responses Minimal efficacy
Interleukin-23 Blockade inhibits Th17 expansion Modest efficacy in phase 3 studies
Interferon-γ Blockade inhibits cytokine production and class II MHC antigen induction Ineffective
Lymphotoxin alpha Blockade inhibits lymphoid organ architecture Ineffective
RANKL Blockade inhibits osteoclast differentiation and systemic and articular bone loss Effective in phase 2 trials for erosions but not
clinical synovitis
Signal transduction
p38 MAP kinase Blockade interrupts cytokine networks Modest, transient efficacy
Bruton’s tyrosine kinase Blockade inhibits B-cell receptor signaling and B-cell activation Marginal-to-modest efficacy, depending on the
compound
PI3 kinase γ or δ Blockade inhibits cell proliferation, survival, and migration to synovium Mixed efficacy
Syk tyrosine kinase Blockade inhibits activation of T cells, B cells, macrophages, and FLS Modest efficacy
Cell-targeting
Cadherin-11 Depletes FLS Ineffective
Rheumatoid Arthritis

CD4 Depletes CD4+ T cells Ineffective

CD52 Depletes CD4+ T cells and several other immune-cell lineages Ineffective

The New England Journal of Medicine

n engl j med 388;6  nejm.org  February 9, 2023
CD5 Depletes certain T lymphocytes and mantle-zone lymphocytes Ineffective
Cell recruitment
ICAM-1 Blocks ICAM–αLβ2 integrin interactions, inhibiting cell recruitment to synovium Ineffective
Multiple chemokines and chemokine Block cell recruitment to synovium Limited or no efficacy for antichemokine anti­

Copyright © 2023 Massachusetts Medical Society. All rights reserved.

receptors bodies and small-molecule chemokine
­receptor inhibitors

* DMARD denotes disease-modifying antirheumatic drug, FLS fibroblast-like synoviocytes, GM-CSF granulocyte–macrophage colony-stimulating factor, ICAM-1 intercellular adhesion
molecule 1, interleukin-1Ra interleukin-1 receptor antagonist, JAK–STAT Janus kinase–signal transducer and activator of transcription, MAP mitogen-activated protein, MHC major his-
tocompatibility complex, PI3 phosphatidylinositol 3, Th1 type 1 helper T cell, Th17 type 17 helper T cell, and TLR toll-like receptor.

Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
537
 The n e w e ng l a n d j o u r na l
cartilage, ligaments, and tendons is destroyed by
proteinases that are produced by synovial cells,
especially fibroblasts, and by chondrocytes them-
selves. The inflammatory cytokine milieu —
most notably, interleukin-1β and TNF — directly
activates these cells to produce matrix metallo-
proteinases, including collagenases, stromelysins
and gelatinases, and ADAMTS5, which contrib-
ute to cartilage and joint destruction.67
Bone destruction requires the action of osteo-
clasts, which differentiate through the combined
actions of the receptor activator of RANKL,68
and proinflammatory cytokines, especially TNF
and interleukin-6.69 The most important source
of RANKL, promoting bone loss in rheumatoid
arthritis, is synovial fibroblasts,70 but certain
T‑cell and B-cell subsets also produce RANKL.
Inflamed synovial tissues and pannus bring in-
flammatory cytokines and RANKL-expressing
cells to the bone microenvironment, inducing
osteoclastogenesis. Osteoclasts attach to bone,
forming an acidic environment that leaches min-
eral from bone, and produce enzymes, including
cathepsin K, that degrade the bone matrix. In
addition, inhibitors of the Wnt signaling path-
way prevent osteoblast differentiation and bone
repair.71,72 RANKL and proinflammatory cyto-
kines enter the circulation, promoting systemic
bone loss and osteoporosis and increasing the
risk of fractures. Systemic bone loss begins in
the preclinical phase of rheumatoid arthritis,
since ACPAs can directly promote osteoclasto-
genesis by triggering Fc receptor activation and
cytokine release from macrophages and acti-
vating osteoclasts.73 Improved therapies and an
aggressive approach to controlling inflamma-
tion in patients have reduced the severity of ar-
ticular and systemic bone loss in patients with
rheumatoid arthritis.74
Ther a peu t ic C onsider at ions
a nd A pproache s
Therapeutic approaches and outcomes in pa-
tients with rheumatoid arthritis have improved
dramatically over the past three decades with
the advent of targeted therapies (Table 2 and
Fig. 3). Early diagnosis and intervention with a
disease-modifying antirheumatic drug (DMARD)
remain the cornerstone of treatment to control
inflammation, prevent joint and organ damage,
and reduce the risk of death.75 Limiting the use
538 n engl j med 388;6  nejm.org  February 9, 2023
The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
of potentially toxic medications such as nonste-
roidal antiinflammatory drugs, glucocorticoids,
and opioids is also an important focus. The use
of composite disease activity measures, often
called “treat to target,” is a critical component
of the treatment strategy.76 Determining the or-
der in which drugs are used is not as important
as selecting one of the many outcome measures
for disease activity that can be incorporated into
the clinical workflow and changing or adding
therapeutic agents as needed to achieve the tar-
get of low disease activity or remission.77 As
successful new therapeutic agents have been
introduced, guidelines for treatment and evi-
denced-based approaches have been updated and
are available to guide clinicians.78,79
Despite the advent of new therapies that tar-
get a wide array of mechanisms, the mainstay
for the initial treatment of rheumatoid arthritis
remains low-dose methotrexate, and 25 to 40%
of patients have substantial improvement with
methotrexate alone.80 Inadequate responses to
methotrexate require the addition of another
agent, typically a biologic agent or a Janus kinase
(JAK) inhibitor. Methotrexate improves the clin-
ical response to many targeted agents when
used in combination therapy. Combinations of
traditional agents (triple therapy) with metho-
trexate, sulfasalazine, and hydroxychloroquine
can achieve adequate responses,81 but adherence
to the regimen may be challenging. Most drug-
specific toxic effects have been well described,
such as bone marrow suppression and liver en-
zyme abnormalities (e.g., with methotrexate and
leflunomide) or thrombosis and an increase in
cardiovascular events (e.g., with the JAK inhibi-
tor tofacitinib, as compared with an anti-TNF
agent).82 Effective drugs typically suppress host
defenses and are associated with low rates of
serious infections (typically ≤1%).83 Glucocorti-
coids are associated with a dose-dependent risk
of serious infection84 and contribute to fracture
and other complications over time.
Many targeted agents evaluated in clinical
trials have limited or no efficacy or have not
been approved for clinical use (Table 2), but we
have learned much from these trials. The site of
action of available therapeutic agents in the con-
text of pathogenesis is shown in Figure 3. Treat-
ment should not simply suppress inflammation
but must also address the specific pathogenic
pathway (or pathways) in an individual patient’s
 Rheumatoid Arthritis

Fc
Immune receptor
complexes
Interleukin-6,
GM-CSF
Cytokines promote
Bone differentiation to
marrow-derived osteoclasts
monocyte or Interleukin-1,
macrophage interleukin-6, Interferon-γ,
interleukin-15, interleukin-17, Interleukin-1,
interleukin-18,TNF TNF TNF
Monocyte
MHC TNF Interleukin-6 or osteoclast
Autoantibodies CD80 or
class II inhibitors inhibitors precursor
including CD86
rheumatoid TCR
factor CD28 Interleukin-6

T cell T cell Interferon-γ, RANK

help interleukin-17, RANKL
TNF
CD40L
Cytokines JAK inhibitors
CD40
Interleukin-6, Baricitinib Fibroblast
Plasma GM-CSF Tofacitinib
B cell Costimulation Upadacitinib
cell
Osteoclast
B–cell Costimulation differentiation
depletion inhibitor
Dendritic GM-CSF and fusion
Rituximab cell
Abatacept
Interferon-α,
interleukin-12, Osteoclast
interleukin-23,
Anti–modified TGFβ
protein Modified
autoantibodies proteins
Proteases

Neutrophil
Bone erosion and
cartilage loss

Figure 3. Synovial Cellular Interactions, Cytokine Networks, and Sites of Action of Current Therapeutic Agents.
Cell–cell interactions within the synovium are critical components of RA pathogenesis. Red boxes show effective therapeutic agents.
Several cell types (dendritic cells, macrophages, and B cells) can present antigens to T cells, including modified (e.g., citrullinated) pro-
teins, to activate these cells and to induce their differentiation. This results in the production of cytokines that, in turn, activate other,
neighboring cells, including monocytes, macrophages, and synovial fibroblasts, to produce additional proinflammatory cytokines and
factors. Neutrophil extracellular traps in the lungs form a scaffold for citrullinated proteins and amplify immune responses that can gener-
ate ACPAs. Activated B cells differentiate into plasma cells that produce ACPAs and other autoantibodies. RANKL is produced by synovial
fibroblasts but also by certain T- and B-cell subsets, inducing the differentiation of monocytes into bone-resorbing osteoclasts. Osteo-
clastic degradation of bone leads to the joint erosions seen in patients with RA, and proteases induced by inflammatory cytokines lead
to cartilage loss and radiographic narrowing of joint spaces. Interleukin-6 inhibitors include tocilizumab and sarilumab. Tumor necrosis
factor (TNF) inhibitors include adalimumab, certolizumab pegol, etanercept, golimumab, and infliximab. Rituximab is an anti-CD20,
B-cell–depleting agent. Abatacept inhibits T-cell costimulation. Janus kinase (JAK) inhibitors include baricitinib, tofacitinib, and upadaci-
tinib. These are positioned between cells producing cytokines including interleukin-6 and inhibit the JAK–STAT (Janus kinase–signal
transducers and activators of transcription) pathway. GM-CSF denotes granulocyte–macrophage colony-stimulating factor, MHC major
histocompatibility complex, TCR T-cell receptor, TGF transforming growth factor.

disease. Methods of stratifying patients to iden-
tify those who are likely to have a response to a
particular treatment, thus improving drug selec-
tion, are being investigated. To date, however,
no combination of biomarkers (e.g., cytokine
levels in the blood), tissue analyses (e.g., histo-
pathological patterns or transcriptome assess-
ments), or genetic markers (e.g., single-nucleo-
tide polymorphisms) has been shown to improve
decision making in clinical practice. Rational

n engl j med 388;6 nejm.org February 9, 2023 539

The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

drug selection remains a critical unmet need,
and the development of alternative taxonomies
based on pathogenesis will be essential.
Disease mechanisms may also vary with the
stage of disease. Epigenetic marks of synovial
fibroblasts in early rheumatoid arthritis are
quite different from those in later stages of the
disease, and the specificities of ACPAs evolve
over time.85 Thus, individualized therapy might
require adjustments for mechanisms that vary as
the disease progresses. This concept is support-
ed by the observation that early disease is more
responsive to therapy than is late disease86 and
highlights the importance of early control of
inflammation.
Clinical remission is the goal of therapy but
is not realized in most patients with rheumatoid
arthritis. Tapering or even discontinuing thera-
pies in patients with complete responses can be
achieved in the short term,87 but unfortunately,
the disease typically recurs. In patients with
early rheumatoid arthritis and low disease activ-
ity, clinical predictors of disease flare after
DMARD discontinuation include measures of
the patient’s functioning and measures of bone
erosion on magnetic resonance imaging.88,89
Rates of disease flare may also differ on the
basis of autoantibody status, as well as the dura-
tion of remission once DMARDs are tapered.90
In a study involving patients with well-controlled
seropositive disease in whom anti-TNF agents
and methotrexate were tapered and stopped, the
cumulative flare rate at 2 years was 61%, and
only 15% of patients had a drug-free remission.91
In addition, therapeutic responses may not be
recaptured when a treatment is reinitiated. Dis-
ease recurrence is likely because most current
therapeutic agents target downstream inflam-
matory mediators rather than resetting the im-
mune system or inducing pathways that resolve
inflammation.
As we increase our understanding of the im-
munologic continuum from a healthy immune
system to preclinical rheumatoid arthritis to
early and chronic disease, new opportunities for
individualized interventions that treat or prevent
disease should emerge. Interceding at the earli-
est time points to prevent disease will perhaps
be as important as identifying new targets for
long-standing rheumatoid arthritis. New classi-
fication criteria are needed to harmonize data
from clinical trials and observational studies
involving at-risk persons. At the same time,
analyses of multiple streams of genomic, pro-
teomic, metabolomic, and epigenomic data are
likely to identify new therapeutic targets and
enable clinicians to select the agent that will
work best in an individual patient.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.

References
1. Ajeganova S, Huizinga TW. Rheuma-
toid arthritis: seronegative and seroposi-
tive RA: alike but different? Nat Rev Rheu-
matol 2015;​11:​8-9.
2. Nielen MM, van Schaardenburg D,
Reesink HW, et al. Specific autoantibod-
ies precede the symptoms of rheumatoid
arthritis: a study of serial measurements
in blood donors. Arthritis Rheum 2004;​
50:​380-6.
3. Sokolove J, Bromberg R, Deane KD,
et al. Autoantibody epitope spreading in
the pre-clinical phase predicts progression
to rheumatoid arthritis. PLoS One 2012;​
7(5):​e35296.
4. Toes RE, Huizinga TJW. Update on
autoantibodies to modified proteins. Curr
Opin Rheumatol 2015;​27:​262-7.
5. Cutolo M, Straub RH. Sex steroids and
autoimmune rheumatic diseases: state of
the art. Nat Rev Rheumatol 2020;​16:​628-
44.
6. Costenbader KH, Karlson EW. Epstein-
Barr virus and rheumatoid arthritis: is
there a link? Arthritis Res Ther 2006;​8:​
204.
7. Möller B, Kollert F, Sculean A, Villiger
PM. Infection triggers in peridontitis and
the gut in rheumatoid arthritis (RA):
a complex story about association and
causality. Front Immunol 2020;​11:​1108.
8. Klareskog L, Padyukov L, Lorentzen J,
Alfredsson L. Mechanisms of disease:
genetic susceptibility and environmental
triggers in the development of rheuma-
toid arthritis. Nat Clin Pract Rheumatol
2006;​2:​425-33.
9. Gan RW, Bemis EA, Demoruelle MK,
et al. The association between omega-3
fatty acid biomarkers and inflammatory
arthritis in an anti-citrullinated protein
antibody positive population. Rheumatol-
ogy (Oxford) 2017;​56:​2229-36.
10. Aletaha D, Neogi T, Silman AJ, et al.
2010 Rheumatoid arthritis classification
criteria: an American College of Rheuma-
tology/European League Against Rheu-
matism collaborative initiative. Arthritis
Rheum 2010;​62:​2569-81.
11. Ishigaki K, Lagattuta KA, Luo Y,
James EA, Buckner JH, Raychaudhuri S.
HLA autoimmune risk alleles restrict the
hypervariable region of T cell receptors.
Nat Genet 2022;​54:​393-402.
12. Raychaudhuri S, Sandor C, Stahl EA,
et al. Five amino acids in three HLA pro-
teins explain most of the association be-
tween MHC and seropositive rheumatoid
arthritis. Nat Genet 2012;​44:​291-6.
13. Viatte S, Plant D, Han B, et al. Asso-
ciation of HLA-DRB1 haplotypes with
rheumatoid arthritis severity, mortality,
and treatment response. JAMA 2015;​313:​
1645-56.
14. Hirose W, Harigai M, Amano K, et al.
Impact of the HLA-DRB1 shared epitope
on responses to treatment with tofaci-
tinib or abatacept in patients with rheu-
matoid arthritis. Arthritis Res Ther 2021;​
23:​228.
15. Lee AT, Li W, Liew A, et al. The
PTPN22 R620W polymorphism associates
with RF positive rheumatoid arthritis in
a dose-dependent manner but not with
HLA-SE status. Genes Immun 2005;​ 6:​
129-33.
16. Svendsen AJ, Gervin K, Lyle R, et al.
Differentially methylated DNA regions in

540 n engl j med 388;6  nejm.org  February 9, 2023

The New England Journal of Medicine

Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.

monozygotic twin pairs discordant for
rheumatoid arthritis: an epigenome-wide
study. Front Immunol 2016;​7:​510.
17. Shao X, Hudson M, Colmegna I, et al.
Rheumatoid arthritis-relevant DNA methyl-
ation changes identified in ACPA-positive
asymptomatic individuals using methylome
capture sequencing. Clin Epigenetics 2019;​
11:​110.
18. Ai R, Boyle DL, Wang W, Firestein GS.
Distinct DNA methylation patterns of
rheumatoid arthritis peripheral blood and
synovial tissue T cells. ACR Open Rheu-
matol 2021;​3:​127-32.
19. Klareskog L, Malmström V, Lundberg
K, Padyukov L, Alfredsson L. Smoking,
citrullination and genetic variability in
the immunopathogenesis of rheumatoid
arthritis. Semin Immunol 2011;​23:​92-8.
20. Liu X, Tedeschi SK, Barbhaiya M, et al.
Impact and timing of smoking cessation
on reducing risk of rheumatoid arthritis
among women in the Nurses’ Health
Studies. Arthritis Care Res (Hoboken)
2019;​71:​914-24.
21. Makrygiannakis D, Hermansson M,
Ulfgren AK, et al. Smoking increases pep-
tidylarginine deiminase 2 enzyme expres-
sion in human lungs and increases citrul-
lination in BAL cells. Ann Rheum Dis
2008;​67:​1488-92.
22. Demoruelle MK, Harrall KK, Ho L,
et al. Anti-citrullinated protein antibod-
ies are associated with neutrophil extra-
cellular traps in the sputum in relatives of
rheumatoid arthritis patients. Arthritis
Rheumatol 2017;​69:​1165-75.
23. Corsiero E, Bombardieri M, Carlotti
E, et al. Single cell cloning and recombi-
nant monoclonal antibodies generation
from RA synovial B cells reveal frequent
targeting of citrullinated histones of NETs.
Ann Rheum Dis 2016;​75:​1866-75.
24. Hill JA, Southwood S, Sette A, Jevni-
kar AM, Bell DA, Cairns E. Cutting edge:
the conversion of arginine to citrulline
allows for a high-affinity peptide interac-
tion with the rheumatoid arthritis-associ-
ated HLA-DRB1*0401 MHC class II mol-
ecule. J Immunol 2003;​171:​538-41.
25. Holers VM, Demoruelle MK, Kuhn
KA, et al. Rheumatoid arthritis and the
mucosal origins hypothesis: protection
turns to destruction. Nat Rev Rheumatol
2018;​14:​542-57.
26. Kongpachith S, Lingampalli N, Ju C-H,
et al. Affinity maturation of the anti-
citrullinated protein antibody paratope
drives epitope spreading and polyreactiv-
ity in rheumatoid arthritis. Arthritis
Rheumatol 2019;​71:​507-17.
27. Kissel T, Ge C, Hafkenscheid L, et al.
Surface Ig variable domain glycosylation
affects autoantigen binding and acts as
threshold for human autoreactive B cell
activation. Sci Adv 2022;​8(6):​eabm1759.
28. de Hair MJ, van de Sande MGH, Ram-
wadhdoebe TH, et al. Features of the
synovium of individuals at risk of devel-
n engl j med 388;6  nejm.org  February 9, 2023
The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
Rheumatoid Arthritis
oping rheumatoid arthritis: implications
for understanding preclinical rheumatoid
arthritis. Arthritis Rheumatol 2014;​ 66:​
513-22.
29. Kuhn KA, Kulik L, Tomooka B, et al.
Antibodies against citrullinated proteins
enhance tissue injury in experimental
autoimmune arthritis. J Clin Invest 2006;​
116:​961-73.
30. Deane KD, Holers VM. Rheumatoid
arthritis pathogenesis, prediction, and
prevention: an emerging paradigm shift.
Arthritis Rheumatol 2021;​73:​181-93.
31. van Boheemen L, Turk S, Beers-Tas
MV, et al. Atorvastatin is unlikely to pre-
vent rheumatoid arthritis in high risk in-
dividuals: results from the prematurely
stopped STAtins to Prevent Rheumatoid
Arthritis (STAPRA) trial. RMD Open
2021;​7:​e001591.
32. Gerlag DM, Safy M, Maijer KI, et al.
Effects of B-cell directed therapy on the
preclinical stage of rheumatoid arthritis:
the PRAIRI study. Ann Rheum Dis 2019;​
78:​179-85.
33. Krijbolder DI, Verstappen M, van Dijk
BT, et al. Intervention with methotrexate
in patients with arthralgia at risk of rheu-
matoid arthritis to reduce the develop-
ment of persistent arthritis and its disease
burden (TREAT EARLIER): a randomised,
double-blind, placebo-controlled, proof-
of-concept trial. Lancet 2022;​400:​283-94.
34. Firestein GS, Yeo M, Zvaifler NJ.
Apoptosis in rheumatoid arthritis
synovium. J Clin Invest 1995;​96:​1631-8.
35. Lewis MJ, Barnes MR, Blighe K, et al.
Molecular portraits of early rheumatoid
arthritis identify clinical and treatment
response phenotypes. Cell Rep 2019;​28(9):​
2455-2470.e5.
36. Ai R, Hammaker D, Boyle DL, et al.
Joint-specific DNA methylation and tran-
scriptome signatures in rheumatoid ar-
thritis identify distinct pathogenic pro-
cesses. Nat Commun 2016;​7:​11849.
37. Frank-Bertoncelj M, Trenkmann M,
Klein K, et al. Epigenetically-driven ana-
tomical diversity of synovial fibroblasts
guides joint-specific fibroblast functions.
Nat Commun 2017;​8:​14852.
38. Zhang, F, Jonsson, AH, Nathan, A,
et al. Cellular deconstruction of inflamed
synovium defines diverse inflammatory
phenotypes in rheumatoid arthritis. Feb-
ruary 28, 2022 (https://www​.­biorxiv​.­org/​
­content/​­10​.­1101/​­2022​.­02​.­25​.­481990v1).
preprint.
39. Humby F, Durez P, Buch MH, et al.
Rituximab versus tocilizumab in anti-
TNF inadequate responder patients with
rheumatoid arthritis (R4RA): 16-week
outcomes of a stratified, biopsy-driven,
multicentre, open-label, phase 4 random­
ised controlled trial. Lancet 2021;​397:​305-
17.
40. Rivellese F, Surace AEA, Goldmann K,
et al. Rituximab versus tocilizumab in
rheumatoid arthritis: synovial biopsy-
based biomarker analysis of the phase 4
R4RA randomized trial. Nat Med 2022;​28:​
1256-68.
41. Ainsworth RI, Hammaker D, Nygaard
G, et al. Systems-biology analysis of rheu-
matoid arthritis patients reveals key regu-
lators and enables mechanism-based
stratification. Nat Commun 2022;​ 13:​
6221.
42. Zhang F, Wei K, Slowikowski K, et al.
Defining inflammatory cell states in
rheumatoid arthritis joint synovial tissues
by integrating single-cell transcriptomics
and mass cytometry. Nat Immunol 2019;​
20:​928-42.
43. Rao DA, Gurish MF, Marshall JL, et al.
Pathologically expanded peripheral T help-
er cell subset drives B cells in rheumatoid
arthritis. Nature 2017;​542:​110-4.
44. Jonsson AH, Zhang F, Dunlap G, et al.
Granzyme K+ CD8 T cells form a core
population in inflamed human tissue. Sci
Transl Med 2022;​14(649):​eabo0686.
45. Meednu N, Rangel-Moreno J, Zhang F,
et al. Dynamic spectrum of ectopic lym-
phoid B cell activation and hypermutation
in the RA synovium characterized by
NR4A nuclear receptor expression. Cell
Rep 2022;​39:​110766.
46. Alivernini S, MacDonald L, Elmes-
mari A, et al. Distinct synovial tissue mac-
rophage subsets regulate inflammation
and remission in rheumatoid arthritis.
Nat Med 2020;​26:​1295-306.
47. Cai B, Kasikara C, Doran AC, Ramak-
rishnan R, Birge RB, Tabas I. MerTK sig-
naling in macrophages promotes the
synthesis of inf lammation resolution
mediators by suppressing CaMKII activity.
Sci Signal 2018;​11:​eaar3721.
48. Culemann S, Grüneboom A, Nicolás-
Ávila JA, et al. Locally renewing resident
synovial macrophages provide a protec-
tive barrier for the joint. Nature 2019;​572:​
670-5.
49. Faas M, Ipseiz N, Ackermann J, et al.
IL-33-induced metabolic reprogramming
controls the differentiation of alterna-
tively activated macrophages and the res-
olution of inflammation. Immunity 2021;​
54(11):​2531-2546.e5.
50. Kuo D, Ding J, Cohn IS, et al. HBEGF+
macrophages in rheumatoid arthritis in-
duce fibroblast invasiveness. Sci Transl
Med 2019;​11(491):​eaau8587.
51. Wei K, Korsunsky I, Marshall JL, et al.
Notch signalling drives synovial fibro-
blast identity and arthritis pathology. Na-
ture 2020;​582:​259-64.
52. Mizoguchi F, Slowikowski K, Wei K,
et al. Functionally distinct disease-associ-
ated fibroblast subsets in rheumatoid ar-
thritis. Nat Commun 2018;​9:​789.
53. Yan M, Komatsu N, Muro R, et al.
ETS1 governs pathological tissue-remod-
eling programs in disease-associated fi-
broblasts. Nat Immunol 2022;​23:​1330-41.
54. Orange DE, Yao V, Sawicka K, et al.
RNA identification of PRIME cells pre-
541

dicting rheumatoid arthritis flares. N Engl
J Med 2020;​383:​218-28.
55. Corsiero E, Pratesi F, Prediletto E,
Bombardieri M, Migliorini P. NETosis as
source of autoantigens in rheumatoid ar-
thritis. Front Immunol 2016;​7:​485.
56. Wehr P, Purvis H, Law SC, Thomas R.
Dendritic cells, T cells and their interac-
tion in rheumatoid arthritis. Clin Exp Im-
munol 2019;​196:​12-27.
57. Lefèvre S, Knedla A, Tennie C, et al.
Synovial fibroblasts spread rheumatoid
arthritis to unaffected joints. Nat Med
2009;​15:​1414-20.
58. Weber B, He Z, Yang N, et al. Diver-
gence of cardiovascular biomarkers of
lipids and subclinical myocardial injury
among rheumatoid arthritis patients with
increased inflammation. Arthritis Rheu-
matol 2021;​73:​970-9.
59. Liao KP, Huang J, He Z, et al. Coro-
nary microvascular dysfunction in rheu-
matoid arthritis compared to diabetes
mellitus and association with all-cause
mortality. Arthritis Care Res (Hoboken)
2021;​73:​159-65.
60. Greenberg JD, Kremer JM, Curtis JR,
et al. Tumour necrosis factor antagonist
use and associated risk reduction of car-
diovascular events among patients with
rheumatoid arthritis. Ann Rheum Dis
2011;​70:​576-82.
61. Baecklund E, Iliadou A, Askling J, et al.
Association of chronic inflammation, not
its treatment, with increased lymphoma
risk in rheumatoid arthritis. Arthritis
Rheum 2006;​54:​692-701.
62. Nerurkar L, Siebert S, McInnes IB, Ca-
vanagh J. Rheumatoid arthritis and de-
pression: an inflammatory perspective.
Lancet Psychiatry 2019;​6:​164-73.
63. Iyer P, Lee YC. Why it hurts: the mech-
anisms of pain in rheumatoid arthritis.
Rheum Dis Clin North Am 2021;​47:​229-
44.
64. Druce KL, Basu N. Predictors of fa-
tigue in rheumatoid arthritis. Rheumatol-
ogy (Oxford) 2019;​58:​Suppl 5:​v29-v34.
65. Dong H, Julien PJ, Demoruelle MK,
Deane KD, Weisman MH. Interstitial lung
abnormalities in patients with early rheu-
matoid arthritis: a pilot study evaluating
prevalence and progression. Eur J Rheu-
matol 2018;​6:​193-8.
66. Juge PA, Lee JS, Ebstein E, et al.
MUC5B Promoter variant and rheumatoid
arthritis with interstitial lung disease.
N Engl J Med 2018;​379:​2209-19.
67. Gravallese EM, Darling JM, Ladd AL,
Katz JN, Glimcher LH. In situ hybridiza-
tion studies of stromelysin and collage-
nase messenger RNA expression in rheu-
matoid synovium. Arthritis Rheum 1991;​
34:​1076-84.
68. Gravallese EM, Manning C, Tsay A,
et al. Synovial tissue in rheumatoid ar-
542 n engl j med 388;6  nejm.org  February 9, 2023
The New England Journal of Medicine
Downloaded from nejm.org at FIOCRUZ on February 9, 2023. For personal use only. No other uses without permission.
Copyright © 2023 Massachusetts Medical Society. All rights reserved.
Rheumatoid Arthritis
thritis is a source of osteoclast differen-
tiation factor. Arthritis Rheum 2000;​43:​
250-8.
69. O’Brien W, Fissel BM, Maeda Y, et al.
RANK-independent osteoclast formation
and bone erosion in inflammatory arthri-
tis. Arthritis Rheumatol 2016;​ 68:​2889-
900.
70. Danks L, Komatsu N, Guerrini MM,
et al. RANKL expressed on synovial fi-
broblasts is primarily responsible for
bone erosions during joint inflammation.
Ann Rheum Dis 2016;​75:​1187-95.
71. Walsh NC, Reinwald S, Manning CA,
et al. Osteoblast function is compromised
at sites of focal bone erosion in inflam-
matory arthritis. J Bone Miner Res 2009;​
24:​1572-85.
72. Diarra D, Stolina M, Polzer K, et al.
Dickkopf-1 is a master regulator of joint
remodeling. Nat Med 2007;​13:​156-63.
73. Harre U, Georgess D, Bang H, et al.
Induction of osteoclastogenesis and bone
loss by human autoantibodies against
citrullinated vimentin. J Clin Invest 2012;​
122:​1791-802.
74. Lindner L, Callhoff J, Alten R, et al.
Osteoporosis in patients with rheumatoid
arthritis: trends in the German National
Database 2007-2017. Rheumatol Int 2020;​
40:​2005-12.
75. Combe B, Landewe R, Daien CI, et al.
2016 update of the EULAR recommenda-
tions for the management of early arthri-
tis. Ann Rheum Dis 2017;​76:​948-59.
76. Smolen JS, Breedveld FC, Burmester
GR, et al. Treating rheumatoid arthritis to
target: 2014 update of the recommenda-
tions of an international task force. Ann
Rheum Dis 2016;​75:​3-15.
77. Grigor C, Capell H, Stirling A, et al.
Effect of a treatment strategy of tight con-
trol for rheumatoid arthritis (the TICORA
study): a single-blind randomised con-
trolled trial. Lancet 2004;​364:​263-9.
78. Fraenkel L, Bathon JM, England BR,
et al. 2021 American College of Rheuma-
tology guideline for the treatment of
rheumatoid arthritis. Arthritis Rheuma-
tol 2021;​73:​1108-23.
79. Smolen JS, Landewé R, Breedveld FC,
et al. EULAR recommendations for the
management of rheumatoid arthritis with
synthetic and biological disease-modify-
ing antirheumatic drugs: 2013 update.
Ann Rheum Dis 2014;​73:​492-509.
80. Emery P, Bingham CO III, Burmester
GR, et al. Certolizumab pegol in combi-
nation with dose-optimised methotrexate
in DMARD-naïve patients with early, active
rheumatoid arthritis with poor prognos-
tic factors: 1-year results from C-EARLY, a
randomised, double-blind, placebo-con-
trolled phase III study. Ann Rheum Dis
2017;​76:​96-104.
81. O’Dell JR, Mikuls TR, Taylor TH, et al.
Therapies for active rheumatoid arthritis
after methotrexate failure. N Engl J Med
2013;​369:​307-18.
82. Ytterberg SR, Bhatt DL, Mikuls TR,
et al. Cardiovascular and cancer risk
with tofacitinib in rheumatoid arthritis.
N Engl J Med 2022;​386:​316-26.
83. Lortholary O, Fernandez-Ruiz M,
Baddley JW, Manuel O, Mariette X, Win-
throp KL. Infectious complications of
rheumatoid arthritis and psoriatic arthri-
tis during targeted and biological thera-
pies: a viewpoint in 2020. Ann Rheum Dis
2020;​79:​1532-43.
84. George MD, Baker JF, Winthrop K,
et al. Risk for serious infection with low-
dose glucocorticoids in patients with
rheumatoid arthritis: a cohort study. Ann
Intern Med 2020;​173:​870-8.
85. Ai R, Whitaker JW, Boyle DL, et al.
DNA methylome signature in synovio-
cytes from patients with early rheumatoid
arthritis compared to synoviocytes from
patients with longstanding rheumatoid
arthritis. Arthritis Rheumatol 2015;​ 67:​
1978-80.
86. Emery P, Solem C, Majer I, Cappelleri
JC, Tarallo M. A European chart review
study on early rheumatoid arthritis treat-
ment patterns, clinical outcomes, and
healthcare utilization. Rheumatol Int 2015;​
35:​1837-49.
87. Lillegraven S, Paulshus Sundlisæter N,
Aga A-B, et al. Effect of half-dose vs stable-
dose conventional synthetic disease-modi-
fying antirheumatic drugs on disease flares
in patients with rheumatoid arthritis in
remission: the ARCTIC REWIND random­
ized clinical trial. JAMA 2021;​325:​1755-
64.
88. Ahmad HA, Baker JF, Conaghan PG,
et al. Prediction of flare following remis-
sion and treatment withdrawal in early
rheumatoid arthritis: post hoc analysis of
a phase IIIb trial with abatacept. Arthritis
Res Ther 2022;​24:​47.
89. Rayner F, Anderson AE, Baker KF,
et al. BIOlogical Factors that Limit sus-
tAined Remission in rhEumatoid arthritis
(the BIO-FLARE study): protocol for a
non-randomised longitudinal cohort study.
BMC Rheumatol 2021;​5:​22.
90. Figueiredo CP, Bang H, Cobra JF, et al.
Antimodified protein antibody response
pattern influences the risk for disease re-
lapse in patients with rheumatoid arthri-
tis tapering disease modifying antirheu-
matic drugs. Ann Rheum Dis 2017;​ 76:​
399-407.
91. van Mulligen E, Weel AE, Hazes JM,
van der Helm-van Mil A, de Jong PHP. Ta-
pering towards DMARD-free remission in
established rheumatoid arthritis: 2-year
results of the TARA trial. Ann Rheum Dis
2020;​79:​1174-81.
Copyright © 2023 Massachusetts Medical Society.